Included Genera:

Apternodus Matthew, 1903.

Distribution and Diagnosis:

As for Apternodus.

Type Species:

Apternodus mediaevus Matthew, 1903

Included Species:

Apternodus gregoryi Schlaikjer, 1933; A. brevirostris Schlaikjer, 1934; A. iliffensis Galbreath, 1953; A. major sp. nov.; A. baladontus sp. nov.; A. dasophylakas sp. nov.

Temporal and Geographic Distribution:

Duchesnean (late middle Eocene) through early Orellan (early Oligocene) throughout western North America.

Diagnosis:

These insectivoran-grade mammals range in size from a large shrew (cf. Crocidura) to Erinaceus. Metacones are absent; talonids are reduced to a single cusp without a basin; and P4/p4 are molariform. The dental formula is 2.1.3.3/3.1.3.3, showing enlarged and procumbent upper and lower anterior incisors. The anterior margin of the coronoid process extends anteriorly to occlude part of m3 from lateral view. The lacrimal foramen is enlarged and faces laterally. The anterior margin of the infraorbital canal is concave, providing attachment space for sizable muscles of the anterior rostrum. A sagittal crest is present, continuous posteriorly with the prominent nuchal crest and forking anteriorly into anterolaterally running crests flanking the posterior margin of the nasal bone. The ethmoid foramen and sinus canal open into the orbitotemporal region via the roof of the sphenorbital fissure. The posterolateral braincase is expanded into lambdoid plates comprised of the squamosal, petromastoid, and occipital bones, all of which are exposed laterally. The mandibular condyle is supported posteriorly by an enlarged entoglenoid process, medial to the postglenoid foramen and anterior to the middle ear. A postglenoid process is absent and the zygomatic arch is incomplete. An ossified auditory bulla is absent; and a piriform fenestra is usually present. The epitympanic recess extends anteriorly into a space bounded ventrally by the entoglenoid process, making the anterior wall of the tympanic cavity concave.

Type Specimen:

AMNH 9601 (left dentary with m1-m2; see fig. 2).

Referred Specimens:

AMNH 9607, left dentary with p3 (see Schlaikjer, 1933: 6, footnote 2); AMNH 76745, posterior braincase with most of left lambdoid plate and petrosal, associated right partial maxilla with I2, C, P2, broken I1, left dentary fragment with p2–3, broken p4, isolated right and left canines, left i3, several other isolated tooth fragments, isolated ?lumbar and caudal vertebrae, right and left proximal humeri with partial right shaft, right distal humerus (fig. 4), metapodial fragments; AMNH 97255, left dentary with p2-m3 (fig. 5); AMNH 97256, right dentary with m1-m3; AMNH 97258, left dentary with p3-m2; USNM, 18966, right dentary with p3-m3; MPUM 0576, right dentary with p4-m2, roots of m3; MPUM 3799, palate with left I1-P3, right C-M1 and alveoli for I1–2, associated with left lambdoid plate (fig. 6); MPUM 6767, left dentary with p3-m3; MPUM 6768, right dentary with p3-m1; MPUM 6855, skull missing anterior rostrum (fig. 7), associated right dentary with p4-m3 (fig. 8); MPUM 6856, left maxillary fragment with canine; MPUM 7820, right dentary with p3-m3; CM 13676, rostrum with left I1-M3, left I2-P2 (original lost, known only from photographs; see fig. 9); CM 31323, right maxilla with C and P2; CM 37455, fragmentary rostrum with right I1-C, left I1, P2, maxillary fragments with partial right M1–2, left P4-M1, associated right and left dentaries with fragmentary right m1-m3, left i1, and roots for p4-m2, isolated lower right i1, c, p2, p3, lower left i3, c, and other fragmentary upper and lower teeth, many disarticulated, fragmentary bits from right and left maxillae and dentaries, isolated cervical and lumbar vertebrae, distal tibia and ulna, partial calcaneus, and phalanx (fig. 10); CM 71569, right maxilla with P2-M2; and CM 71570, right dentary with p3-m2.

Temporal and Geographic Distribution:

Middle to late Chadronian (Late Eocene) of southwestern Montana (Pipestone Springs 29, Canyon Ferry 6, West Easter Lily 13, Highway 10N 22, and Little Pipestone Creek 25). In this section and those that follow, numbers given after localities correspond to those depicted on the map shown in figure 1.

Diagnosis:

The upper and lower canines have elongate, bulbous crowns. The upper canine has two broad roots, anterior and posterior; some specimens (e.g., MPUM 3799; fig. 6) show a small, lingually situated third root. The P2 is slightly enlarged and elongate, with a distinct crest running distally from the main cusp and a slender, lingual, third root extending from the midpoint of the tooth. The i2 is diminutive with a small, poorly defined alveolus that abuts against the much larger i1. The i3, canine, p2, and p3 are also enlarged and bulbous. The p3 shows a distinct posterior cusp and a slight buccal cingulum. The upper molars have distinct protocones. Buccal cingulids are present on the lower molars. Basicranially, the rostral tympanic process of the petrosal exhibits a medial ridge. The posttympanic process extends well ventral to the external auditory meatus.

Remarks:

Particularly noteworthy among the Apternodus mediaevus specimens not previously discussed in the literature are MPUM 3799, an associated rostrum and partial posterior braincase (fig. 6); AMNH 76745, a fragmentary skull with associated postcranial fragments (fig. 4); MPUM 6855, a partial skull (fig. 7) and associated mandible (fig. 8); CM 13676, a rostrum and complete upper dentition (fig. 9; original lost); CM 37455, a fragmentary skull and mandibles associated with several broken postcranial elements (fig. 10); and AMNH 97255, a well-preserved dentary (fig. 5). These specimens provide a fairly complete picture of the skull of A. mediaevus, as well as some information about its skeleton. With the exception of the more gracile upper P2 and I2, the more ventrally pronounced lambdoid plate, the diminutive i2, and the posterior cusp and cingulid on p3, A. mediaevus resembles A. baladontus sp. nov., described below.

The Canyon Ferry specimen preserving associated cranial and postcranial fragments (CM 37455; White, 1954) is included in Apternodus mediaevus based on the morphology of its I2, P2, p3, and enlarged, bulbous upper and lower canines. However, this specimen is larger than others in A. mediaevus; its inclusion increases the size range in this species (see later section on metric variation). The only other Apternodus specimen known from Canyon Ferry (USNM 18914, a maxillary fragment with M1 and alveoli for P3-P4, M2-M3) is smaller and well within the range of variation for A. mediaevus. Nevertheless, considerable variation in size does occur in some modern taxa (see below); more importantly, CM 37455 is anatomically cohesive with specimens in the A. mediaevus hypodigm, and for the present is retained in this species.

One mandible in the AMNH collection from Pipestone Springs (AMNH 76747, a left dentary with an erupting p3 and broken m1-m3) does not fit into existing specific diagnoses, and may represent yet another species from Montana. This specimen differs from A. mediaevus in showing a very gracile coronoid process and dentary. The coronoid does not anteriorly overlap with m3, and the mandibular corpus is thinner and more gracile than those of any other Pipestone Springs specimen, including one with erupting permanent premolars (CM 71570). Some breakage and postmortem damage has made the mandibular condyle of AMNH 76747 shorter mediolaterally than it was in life. Although the trigonids of all three molars of AMNH 76747 are severely damaged and the intact p3 is not yet fully erupted, remnants of its teeth closely resemble those of other Apternodus specimens from Pipestone Springs. In particular, the p3 is bulbous, with a distinct posterior cusp and weak buccal cingulum. Because of these similarities and because AMNH 76747 represents a juvenile and possibly pathologic individual, we refrain from recognizing it as a new species.

Type Specimen:

FMNH UM1690, complete skull (fig. 11) with both mandibles (fig. 12). Damage to FMNH UM1690 is confined mainly to the occiput, the anterior upper incisors, and the coronoid process of the left mandible. Part of the cerebellar cast is exposed by damage to the posterior parts of the parietals. Otherwise, the specimen is very well preserved. The frontispiece of this monograph shows a reconstruction of Apternodus baladontus as it may have appeared in life, based on this specimen.

Referred Specimens:

MPUM 0450, left dentary with broken p3-m2; MPUM 2620, rostrum with left I2-P3, right C-P2; MPUM 2634, rostrum (fig. 13) and mandibles (fig. 14) with nearly complete upper and lower dentition, missing anterior incisors and crown of left m1; MPUM 2645, right maxilla with canine; CM 9552, right dentary with p4-m2, erupting p3; CM 71563, right dentary with p3-m3.

Temporal and Geographic Distribution:

Late Duchesnean (late middle Eocene) to middle Chadronian (late Eocene) of southwestern Montana (McCarty's Mountain 27, Diamond O Ranch 9, and Little Pipestone Creek 25).

Etymology:

Named for the sphere-shaped anterior dentition, based on the Greek words for “ball” and “tooth”.

Diagnosis:

In addition to large, bulbous upper canines, bulbous i3-p3, and prominent molar protocones, A. baladontus shows a large, spherical upper P2 and I2, with P2 showing at least three and sometimes four roots (e.g., MPUM 2645). The upper canine has a prominent, buccally situated third root. The p3 lacks a prominent posterior cusp. A. baladontus is further distinguished from A. mediaevus by the small but distinct i2 which has an alveolus separate from that of i1, small cheek teeth with molars that show weak or absent buccal cingulids, the presence of a bony torus defining the ventrolateral margin of the external auditory meatus, and the ventrally short posttympanic process (fig. 11).

Remarks:

In addition to the type, the other exceptional element of the A. baladontus hypodigm is MPUM 2634, a rostrum (fig. 13) with complete mandibles (fig. 14). This specimen is unusual in that it preserves a suture between the maxilla, palatine, and frontal in the orbital mosaic, and demonstrates that Apternodus possessed a maxilla that extended far into the orbit. Remnants of the orbital sutures of the maxilla are also evident in a specimen of A. mediaevus (MPUM 6855; fig. 7). In other specimens (e.g., AMNH 22466 and MCZ 17685) these sutures are fused, and do not permit estimates of the maxillary contribution to the orbital mosaic (contra McDowell, 1958). Although many of its cranial sutures are untraceable, the A. baladontus type (FMNH UM1690) is otherwise complete enough to warrant particular attention, as follows.

Rostrum:

The nasals, maxillaries, and premaxillaries have completely fused. Above the enlarged upper incisors, a shallow pit can be seen on each premaxillary, just posterior to the anterior end of the bony snout. The dorsomedial boundary of a large muscle scar extends from the medial edge of these pits back to a point above the lacrimal foramen. This scar covers most of the side of the snout and evidently supported powerful musculature for the movement of some sort of proboscis. Similarity to the snouts of Solenodon, Nesophontes, and Centetodon (Lillegraven et al., 1981) is marked. The anterior rim of the orbit is likewise similar to that of Nesophontes, Solenodon, and Centetodon. There is a large lacrimal foramen, from which the nasolacrimal canal runs anteromedially across the roof of the infraorbital canal. The latter is relatively short. A short nubbin over M1 is all that remains of the zygomatic wing of the maxillary bone. In other insectivorans (e.g., most tenrecs and Solenodon), this structure is usually larger and more posteriorly located. Several nutritive foramina open into the infraorbital canal. The sphenopalatine and greater palatine foramina open separately into the lower part of the orbit and are even with the posterior end of the palate.

The palatines extend forward to a point opposite the posterior half of M1. Their combined anterior border takes a straight course across more than half of the palate, although the squamous nature of the contact results in an irregular series of about eight forward-projecting lobes of bone. The posterior border of the palate terminates relatively farther behind the level of M3 than in A. brevirostris, in agreement with the condition in A. gregoryi. The border is swollen at the internal nares and projects somewhat ventrally, but a true torus is not present, nor is there any posterodorsal shelf behind the swelling. The border is instead notched at the midline in this specimen. There are no palatal fenestrae.

At least seven pairs of palatal foramina are present posterior to the level of P2. The most posterior pair is the largest, situated posteromedial to M3, and presumably carried the greater palatine artery. Dorsally, the opening of the greater palatine foramen is ventrolateral to that of the sphenopalatine foramen; the two are not coincident. The greater palatine artery apparently divided into a number of branches that emerged onto the palate at intervals. A small, paired foramen can be seen anteromedial to the opening for the minor palatine artery at the level of M3. Anterolabial to this foramen lies another paired foramen in the maxillopalatine suture opposite the interdental embrasure between M2 and M3; this gives rise to a faint groove that runs forward on the surface of the maxillary close to the lingual roots of the molars. This may be related to a feature called the third supply tract of the palate in Centetodon (McKenna 1960: 139). The next pair of foramina is located at the anterior limits of the palatine bones. Approximately 1 mm from each side of the midline, a large vascular foramen is present. The maxillary is grooved from there to a point opposite P3. The groove from each foramen of the pair divides into two slightly divergent grooves. Between the lingual roots of P2 and P3 are two small foramina on each side. Anteriorly, no additional paired foramina are present except for the large incisive foramina between the incisors.

Orbitotemporal Region and Basicranium:

The ventrally projecting pterygoids are similar to those of several insectivorans, including Microgale and Solenodon. Only a single pair of ventral projections is present; each of these hooks posteriorly opposite the posterior opening of the alisphenoid canal as in Microgale cowani and probably supported a membrane separating the trachea from the rear of the oral cavity, in addition to musculature of the soft palate and auditory tube. Lateral to the ventral border of the pterygoid is a flat surface for attachment of the internal pterygoid muscle.

The sphenorbital fissure and associated foramina in FMNH UM1690 are quite distinctive and well preserved. A massive overhanging wall of bone extends forward as a continuation of the lateral wall of the sphenorbital fissure, terminating at a point approximately even with the posterior border of the palate. In the roof of the resulting vault are several foramina, all of which are hidden from lateral view. The most anterior of these is the ethmoidal foramen, which leads into the anterior cranial fossa and posterior ethmoidal region. Immediately posterior to the ethmoidal foramen are two tiny foramina of uncertain significance. Posterior to these is the large anterior opening of the sinus canal, which in life transmitted the superior ramus of the stapedial artery along the interior of the braincase. Medial to this is a small optic foramen. Immediately posterior to the optic and sinus canal foramina is the much larger opening for the ophthalmic and maxillary divisions of the trigeminal nerve. No separate foramen rotundum is present. Other specimens that preserve the sphenorbital fissure in good condition include some with breakage that permits examination of these foramina from both dorsal and ventral perspectives (e.g., A. brevirostris AMNH 74951). These specimens share the morphology exhibited by FMNH UM1690.

Behind the posterior base of the sphenorbital fissure, just anterior to foramen ovale, is the small posterior opening of the alisphenoid canal. It is about 0.5 mm in diameter and opens directly into the floor of the sphenorbital fissure. The foramen ovale is about 0.7 mm in diameter and faces anteroventrally, medial to the glenoid fossa.

Apternodus baladontus possesses a piriform fenestra, as the squamosal and alisphenoid bones fail to contact the anterolateral edge of the petrosal. The rostral tympanic process of the petrosal (MacPhee, 1981) projects ventrally along the area postulated (erroneously) by McDowell (1958: 169) to contain a shallow groove for the internal carotid artery in A. brevirostris.

The anterior carotid foramen, through which the internal carotid artery presumably entered the braincase, lies in the petrosal-sphenoid suture at the anteromedial corner of the tympanic roof, just medial to the point at which the suture separates to produce the piriform fenestra. Further expansion of the piriform fenestra would engulf the anterior carotid foramen, resulting in passage of the internal carotid artery through the enlarged fenestra, much as in soricids. Slightly anteromedial to the anterior carotid foramen, a small opening in the alisphenoid is evident; this presumably transmitted a vidian branch of the internal carotid artery.

Jaw Joint:

Figure 25 of McDowell (1958: 170) is an excellent illustration of the differences between the entoglenoid and postglenoid processes. McDowell recognized that for posterior support of the mandibular condyle, Solenodon and Nesophontes (along with soricids, moles, and tenrecs) possess a modified entoglenoid process located medial to the postglenoid foramen and anterior to the promontory of the petrosal. These taxa lack a true postglenoid process (see also MacPhee, 1981: 220). Strangely, McDowell did not recognize that this was also the case for Apternodus. As exemplified by the type of A. baladontus, and similar to the condition in Tenrec figured by McDowell (1958: fig. 25C), the mandibular condyle of Apternodus is supported posteriorly by a flange of squamosal that is coincident with the anterior border of the middle ear, grooved by the chorda tympani, and located anteromedial to the postglenoid foramen. Some variation exists in the contribution of the alisphenoid to the entoglenoid process; in Tenrec it is considerable, while in Solenodon and Apternodus it is small.

This contrasts with a true postglenoid process (as illustrated in Leptictis by McDowell 1958: fig. 25A), which is located anterolateral to the postglenoid foramen, and well lateral to the promontory of the petrosal and path of the chorda tympani. McDowell's figure 24 (p. 169) accurately shows the morphology of the entoglenoid process in Apternodus, which can be seen to differ as noted above from the postglenoid process of Leptictis.

Braincase:

The parietals of FMNH UM1690 are quite rugose, a feature in keeping with powerful temporal muscles. The ventral margin of the lambdoid plate does not reach as far ventrally as that of other members of the genus Apternodus. In contrast, other species show a ventral mastoid rim of the lambdoid plate well below the level of the jaw joint.

Hyoid Apparatus:

In the course of preparation of the skull, two hyoid elements were found. These appear to be (1) an ankylosed complex made up of the basihyal and both thyrohyals and (2) either a right ceratohyal (of Sprague, 1944, or epihyal of Allen, 1910) or a right hypohyal (also of Sprague, 1944, or ceratohyal of Allen). If the latter really is a ceratohyal, as seems probable because of its large size relative to the basihyal-thyrohyal complex, then the hypohyal must have been very small, a special similarity to Solenodon among insectivorans in which the hyoid region is known (Sprague, 1944: table l). If the bone is a large hypohyal, then there is little similarity to that of any other insectivoran.

Dentition:

The dental formula of A. baladontus is the same as that of other Apternodus species: 2.1.3.3/3.1.3.3. The few Apternodus specimens that preserve erupting teeth (e.g., AMNH 76747, CM 9552, CM 71570) have three fully erupted molars posteriorly, and are thus consistent with the replacement pattern observed in most other placental mammals. Based on this pattern, we assume that the premaxillary teeth of taxa discussed here are incisors, the anteriormost maxillary tooth is a canine, and the anterior three lower teeth are incisors, followed by a canine.

Accepting these assumptions, there are only two upper incisors in FMNH UM1690, identified here as I1 and I2 based on the observation that in some other insectivoran-grade placentals, I3 is small or absent (e.g., Solenodon and Tenrec). The crown of I1 is not preserved on either side of the skull. The root of the left I1 is 2.0 mm in anteroposterior diameter just within its alveolus. The right and left I1 pitch toward each other and may have met medially, depending on the length of the crown. The tips of the enlarged anterior lower incisors pass forward between the right and left I1 and are accommodated by a gutter in the bone below the nares when the mouth is closed. Lateral motion of the anterior end of the occluded mandible is thereby restricted. I2 has a bulbous crown, and is smaller and more spherical than the elongate, semipremolariform I2 of A. mediaevus. The root is straighter than in CM 13676 or A. brevirostris, but the tooth is tilted medially in the same manner. Wear is confined to a flat facet at the apex of the tooth.

The upper canine is a massive crushing tooth with a bulbous conical crown, supported by three roots. The anterior and posterior roots are massive and diverge slightly when seen from the side. The third, medial root supports the anterolabial corner of the tooth and projects anterodorsally in the wall of the snout, approximately parallel with the anterior root. The crown overhangs the root medially. Wear is confined to a flat surface at the apex. The wear surface tilts slightly medial from the horizontal.

As in most other placentals with three premolars, we consider P1/p1 (or dP1/dp1) of Apternodus to be absent. P2 possesses two main roots and has a third, posterolingual root that is also present in A. mediaevus. The crown is circular and bulbous, larger than P2 of A. brevirostris, but less elongate than P2 of A. mediaevus. Wear is confined to a flat surface at the apex of the crown. P3 is small and triangular, with a weak posterolabial shearing crest. The cingula meet lingually, but a protocone is only faintly indicated. The molars are more gracile, but are otherwise similar to those of A. mediaevus from Pipestone Springs.

The anteriormost lower incisor (i1) is greatly enlarged and procumbent, and contacts the opposite anterior incisor for about half the length of its crown. The root is oval in cross section and, in contrast to Solenodon, is not hollowed out medially. No serrations are present along the upper edge of the labial enamel surface. The second lower incisor is small, but not as diminutive in FMNH UM1690 as it is in A. brevirostris; unlike the latter, the i2 of A. baladontus has a distinct alveolus. As in the upper dentition, i3 to p2 present a series of bulbous, procumbent crowns, presumably adapted for crushing. This curious morphology somewhat resembles an overlapping series of inverted spoons. All are single-rooted.

The lower canine and p2 are similar in appearance and possess bulbous crowns. This contrasts with species of Apternodus common outside of Montana, in which both teeth are premolariform. The ventral side of the anterior part of the canine crown of FMNH UM1690 is hollowed to fit the rear of the crown of i3. No trace of a central high cusp or ridge remains. The second lower premolar of A. brevirostris possesses a crested crown, whereas p2 of A. baladontus does not. The third lower premolar is the first double-rooted lower cheek tooth, and is more premolariform than anterior teeth. Its crown forms a low cone with a very weak posterior crest that runs from the apex to a tiny ridge that lies directly ventral to the paraconid of p4. The first and second lower molars are approximately equal in size; m3 is slightly smaller. Buccal cingulids on the molars of the type (fig. 12) and most referred specimens (e.g., fig. 14) are very slight.

The teeth of FMNH UM1690 seem to retain traces of pigment once present in life. Pigment is present in various soricid teeth as far back as the Oligocene (Patterson and McGrew, 1937: 247) and can also be seen in Solenodon (especially under ultraviolet illumination). The pigment is visible as an opacity blocking the normal yellow or orange radiation emanating from the enamel near the apices of cusps. In the preserved upper dentition of FMNH UM1690, only the left I2 is fully pigmented, but traces of pigment can be seen within the enamel of all teeth, especially M2, distributed near cusp apices. The lower anterior incisors are heavily pigmented on the labial and medial faces, but not on the preserved upper face. Similar pigmentation is present in the type specimen of A. brevirostris, in the upper dentition of the type specimen of A. iliffensis, and in some specimens from Pipestone Springs (e.g., AMNH 97255). It is of course possible that in some cases “pigment” may be an artifact of fossilization, and is in any event obscured in specimens (e.g., MPUM 3799, AMNH 76745, USNM 18966, CM 71569) with jet-black teeth. Although the distribution of pigment in Apternodus teeth resembles that of soricids and probably had a biological cause, we note that many vagaries of preservation and wear heavily influence whether or not pigmentation is observable in any given specimen. We therefore refrain at present from attributing much systematic importance to this character.

Type Specimen:

AMNH 22466, slightly damaged complete skull (fig. 15) with associated lower left dentary, complete except for the crown of i1 and coronoid process (fig. 16). This specimen was originally accessioned into the UW collections (Schlaikjer, 1933: 15) and is probably associated with UW 26, a right dentary with p4-m3.

Referred Specimens:

AMNH 74941, broken skull with left P2-M2; AMNH 74942, broken skull, missing anterior rostrum, associated with left dentary with m2-m3 and unprepared right dentary with p4-m3; AMNH 74946, rostrum with left P3-P4 and broken M1; AMNH 74948, rostrum with right P2-M2, broken M3, left C, P3-M3; AMNH 74949, rostrum with right P2-M1, broken M2, left P2-M2; AMNH 74950, half rostrum with left P3-M3; AMNH 74951, broken skull preserving orbitotemporal regions, left petrosal, fragments of left and right lambdoid plates, and rostrum with left P2, broken P4-M3, right canine, broken P3-M1, M3, associated left dentary with p4-m3 and right dentary with m1-m3, also associated with fragmentary left distal humerus; AMNH 76692, left maxillary fragment with P2-M3; USNM 437460, rostrum (fig. 17) with right and left C-M3, left I2, alveoli and roots for both I1s and right I2, missing buccal margins of right P3-M3; associated with left (fig. 18) and right dentaries with i2-m3 and broken i1s. (For purposes of exhibition, and long before this writing, USNM 437460 was attached to the basicranium of another individual from a different part of the Flagstaff Rim section.)

Temporal and Geographic Distribution:

Chadronian (late Eocene) of central Wyoming (Flagstaff Rim 18, Bates's Hole 3).

Diagnosis:

As the name implies, the rostrum is short, holding a closely packed dentition anterior to P4, usually with overlap between the posterior root of P2 and the anterior root of P3 and without diastemata on either side of the upper canine. The anterior dentition is gracile, with premolariform upper and lower I2-P3. The upper molars have lingually situated protocones that are similar in size to the molar parastyles. The i2 is very small (Schlaikjer, 1934), with an alveolus continuous with that of the larger i1. Both the upper and lower second premolars are much smaller than adjacent teeth; and the p3 shows a prominent buccal cingulum and posterior cusp. The rostral tympanic process of the petrosal is medially flat. The external auditory meatus is ventrally concave. The lacrimal foramen is offset posteriorly from the orbit by the superior margin of the infraorbital canal.

Remarks:

Several rostra with complete or nearly complete dentitions are now known for A. brevirostris (AMNH 74941, 76692, 74950, 74948, 74949, and USNM 437460), most of which were collected by Morris Skinner and colleagues from the Flagstaff Rim area, Wyoming during the 1950s. The type (AMNH 22466) remains the most complete specimen known. Surprisingly, very few specimens have been recovered with associated mandibles. One of these is AMNH 74951, which preserves not only a rostrum and posterior mandibles, but is also associated with a distal humerus, petrosal, and fragmentary lambdoid plate. Several braincases and basicrania are known from Flagstaff Rim, including AMNH 74940, one the few specimens known to preserve an ectotympanic (fig. 19). These are presumably referable to A. brevirostris, but without the dentition it is impossible to be certain. Frustratingly, no specimen unambiguously referable to A. brevirostris preserves the ventral margins of the lambdoid plates. These can nevertheless be inferred to have extended well below the level of the glenoid fossa, based on the size of their breakage scars, which are quite large posterior to the external auditory meatus and presumably extended for several millimeters farther ventrally (as reconstructed by Schlaikjer, 1934).

Although a closely packed I2-P4 toothrow characterizes this species, some variation is present in the size of the P2-P3 diastema. For example, in AMNH 74948 (identifiable as A. brevirostris by virtue of its upper molar protocones, gracile anterior dentition, small canine diastemata, and ridge separating the lacrimal foramen from the orbit) the space between P2 and P3 is small, but the roots of the two teeth do not overlap. The ratio of palate length to width in A. brevirostris is smaller than that of other Apternodus species, but not always with decisive statistical significance (fig. 20). Within the A. brevirostris hypodigm, the rostrum of the type specimen (AMNH 22466) is one of the shortest.

McDowell (1958) published a discussion and four illustrations of Apternodus that contained several mistakes. He based these illustrations primarily on the type of A. brevirostris (AMNH 22466), with missing parts restored from the type of A. gregoryi (MCZ 17685) and with the inferred morphology of one side reconstructed from a mirror image of the other. Both skulls are imperfectly preserved and do not show a number of characters depicted in McDowell's reconstruction.

First, McDowell (1958: fig. 23A) repeated Schlaikjer's error (1933: fig. 3, top; see also section below on A. gregoryi) of running a suture between the parietal and occipital (Schlaikjer's mastoid) bones diagonally across the posterolateral corner of the skull roof. There is no evidence for a suture at this site.

Similarly, there is no evidence to support McDowell's (1958: fig. 23) placement of the squamosal-parietal, fronto-parietal, and fronto-nasal sutures. The dorsal exposures of most dermal braincase elements are completely fused, and leave no indication of any sutures.

McDowell (1958: 170) argued that “from the American Museum skull (AMNH 22476 [sic, actually 22466]), it can be determined that the maxilla forms little of the orbital wall, an important contrast to the Lipotyphla.” In fact, the sutures of the maxilla in the orbital mosaic are completely fused in both AMNH 22466 (fig. 15) and MCZ 17685 (fig. 21), as they are in the majority of other Apternodus skulls, with the two exceptions mentioned above: MPUM 2634 (A. baladontus) and MPUM 6855 (A. mediaevus). In the former, a suture is clearly visible on both sides curving around the anterior border of the sphenopalatine foramen and posterior to the lacrimal region (fig. 13). Similarly, both sides of the skull in MPUM 6855 (fig. 7) show distinct, wavy lines representing the maxillary-frontal suture and indicating that the maxillary incursion into the orbital mosaic of Apternodus is considerable, similar to that of Erinaceus (MacPhee and Novacek, 1993: fig. 3.5).

McDowell's reconstructed cribriform plate (1958: fig. 23) is located too far posteriorly, and is oriented too vertically. A partial, broken skull of A. brevirostris (AMNH 74951) exposes most of the cribriform plate intact, and shows its posteriormost point just anterior to the sella turcica, as in most other mammals. Externally, this corresponds with a point just anterior to the glenoid fossa of the mandible, close to the posteriormost external opening of the sphenorbital fissure. Relative to this point, the cribriform plate in Apternodus extends both anteriorly and dorsally toward the anterior end of the sagittal crest, to a point coincident with a coronal plane shared with the posterior margin of the hard palate.

McDowell's figure 23C incorrectly shows a completely solid tympanic roof, with no piriform fenestra on either side. In fact, the region immediately anterior to both petrosals in AMNH 22466 is perforate, although the large size of these “fenestrae”, particularly on the left side, is due to postmortem damage. Nevertheless, most specimens of Apternodus that preserve the relevant anatomy (e.g., FMNH UM1690, AMNH 74941, USNM 455680, UW 11046) clearly show prominent piriform fenestrae anterior to the petrosal. This was shown in AMNH 22466 by Matthew (1910), who included photographs in his original description of the specimen. The type of A. gregoryi (MCZ 17685; see below) was also incorrectly reconstructed as lacking piriform fenestrae by Schlaikjer (1933: fig. 3) and Scott and Jepsen (1936: plate 1). In some individuals (MPUM 6855 and AMNH 74940) the fenestrae are indeed small or absent; AMNH 74942 appears to show a piriform fenestra on the left side but not on the right, although this is undoubtedly the result of postmortem damage. In general, the tympanic roof anterior to the petrosal is naturally perforate in most specimens of Apternodus, including the type of A. brevirostris.

McDowell (1958: figs. 23C, 24, and p. 168) argued that the internal carotid artery of A. brevirostris runs medial to the middle ear after giving off a stapedial branch that occupies a groove on the promontorium ventral to the vestibular foramen. This pattern, he wrote, was “very different … from that of the Lipotyphla, [and is] more like that of creodonts …” (p. 168). In fact, as discussed elsewhere (Asher, 2000, 2001; Whidden and Asher, 2001) McDowell overestimated the conservatism of cranial blood supply among insectivoran mammals, and drew a misleading contrast between an alleged “lipotyphlan pattern” and a speculative model of arterial circulation presented for Apternodus based on AMNH 22466 and MCZ 17685. In short, although there is a faint arterial groove associated with the proximal stapedial artery in the immediate vicinity of the vestibular foramen in the petrosal of AMNH 22466, there is no indication of a groove for the internal carotid artery. Neither the medial aspect of the petrosal of AMNH 22466, nor its ventral apex, is marked by subtle depressions that in other specimens define the course of the internal carotid artery along the ventrum of the tympanic roof. As noted elsewhere (Whidden and Asher, 2001), other specimens of Apternodus (e.g., AMNH 74942, 76745; USNM 455680) do preserve shallow grooves running between the anterior carotid foramen and ventral apex of the petrosal in a position well lateral to the petrosal-basisphenoid suture. These specimens indicate either that arterial circulation in Apternodus was polymorphic, or more likely, that McDowell's vascular reconstruction (1958: fig. 24) was incorrect. In sum, it is doubtful that the course of the internal carotid artery in Apternodus was located farther medially than that of many insectivorans, such as Microgale (MacPhee, 1981; Asher, 2001).

Type Specimen:

MCZ 17685, complete skull (fig. 21) with left C, P3-M2, right C, P2-M2, broken M3, missing premaxillae and anterior dentition, associated mandibles with left p3-m3, right i1, and p3-m3.

Referred Specimen:

UW 13508, skull (fig. 22) with intact but distorted basicranium, palate with left I1 root, C-M3, roots for right I1–2, plus right C-M3, associated with mandibles (fig. 23) including broken left i1, i3-c, p3-m3, and broken right i1, i3-m3. Also associated with proximal and distal fragments of the femur and os coxae (fig. 24), a fragment of the proximal ulna, five fragmentary caudal vertebrae, and several indeterminate fragments of ribs and/or longbones.

Temporal and Geographic Distribution:

Chadronian (late Eocene) to Orellan (early Oligocene) of eastern Wyoming (Dilts Ranch 10; Torrington 38).

Diagnosis:

The rostrum of Apternodus gregoryi shows small diastemata between P2 and P3 and on either side of the canine. The external auditory meatus is very narrow and defined posteriorly by a ventrally projecting posttympanic process. Anteriorly, this notch is defined by a ventrally elongate lower lip of the entoglenoid process. The rostral tympanic process of the petrosal is flat. The lacrimal foramen of A. gregoryi is posteriorly continuous with the anterior orbit. That is, unlike A. brevirostris, iliffensis, mediaevus, and baladontus (but similar to A. major), the superior margin of the infraorbital canal is set apart from the lacrimal foramen posteriorly, leaving this region without a ridge of bone defining the posterior aspect of the foramen (figs. 21, 22). Protocones on the upper molars are reduced, smaller than their corresponding parastyles. The cingulid and posterior cusp of p3 are reduced.

Remarks:

The only specimen known of A. gregoryi besides the type is UW 13508. This specimen consists of a nearly complete skull (fig. 22) and lower dentition (fig. 23) from both sides, as well some postcranial elements (fig. 24). The skull is quite similar to MCZ 17685, although the ventral lambdoid plates of UW 13508 are broken. The large size of their breakage scars indicates that the lambdoid plates were ventrally extensive. What remains of the external auditory meatus is sharply concave, as in the A. gregoryi type specimen (fig. 21).

Postcranially, the femur and partial os coxae of UW 13508 appear small for an animal with a skull of this size. However, the ratio of anteroposterior skull length to indices of proximal femur and acetabular size (fig. 25) is similar to that of extant groups such as talpids and soricids; hence, we assume that the partial femur, os coxae, and skull represent the same individual. In fact, prior to preparation, the distal femur of UW 13508 was embedded in matrix adjacent to the basicranium, making clear at least that the two specimens were deposited in close proximity, although in an anatomically unexpected position. The dentition of the associated skull is fully erupted, but shows minimal wear. Furthermore, the sphenoccipital synchondrosis of the skull is not completely fused (fig. 22), consistent with its association with the juvenile represented by the partial femur.

The distal femur preserves only the epiphysis. Proximally, the femur also appears to be detached from the epiphysis, although in this case the epiphysis is missing. Perhaps as a consequence, the femoral head lacks any sign of a fovea capitis. If what remains of the proximal femur (missing the epiphysis) accurately reflects the shape of the femoral head, then it is dorsally flattened and not separated from the greater trochanter by a deep dorsal concavity, in contrast to the morphology seen in most other insectivoran-grade taxa (e.g., Echinosorex; see MacPhee, 1994: 131). Extant animals with a similar femoral morphology include Didelphis, Orycteropus, Tenrec (MacPhee, 1994: 130–131), and soricids. To varying degrees, these taxa show arboreal, fossorial, and terrestrial locomotor patterns and do not present a simple locomotor analogue to Apternodus.

Schlaikjer's (1933) study of Apternodus gregoryi was considerably more detailed than that of Matthew (1910), but contained the following points that we cannot confirm.

First, Schlaikjer (1933: 10 and fig. 4) thought that there was some evidence that the mastoid extended onto the occiput to contact the exoccipital and supraoccipital. However, the mastoid-occipital suture actually lies on the lateral-facing lambdoid plate, as noted by Matthew and later by Schlaikjer himself (1934: 6). There is no visible exoccipital-supraoccipital suture on skulls of A. brevirostris or A. gregoryi. The entire occiput is composed of fused parts of the occipital bone.

The squamosal was believed to extend back on the lambdoid plate to what in reality is the mastoid-occipital suture. Schlaikjer's (1933: fig. 3) figure of the lateral view of the skull omitted the squamosal-mastoid suture. The latter was corrected by Schlaikjer (1934) and McDowell (1958: 165, fig. 23).

On the lambdoid plate, Schlaikjer's (1933: fig. 3, top) identification of the lateral exposure of the occipital as a part of the mastoid bone led him to depict a faint “mastoid-parietal” suture on the dorsal view of the skull. A suture between the occipital and the parietals was assumed to run along the lambdoid crests. A similar interpretation was given by McDowell (1958: fig. 23A). However, in available material of A. gregoryi, dorsal exposures of dermal cranial bones are fused and sutures cannot be accurately distinguished.

As noted above, piriform fenestrae were incorrectly shown as absent by Schlaikjer (1933, 1934) in both A. gregoryi and A. brevirostris. The type of A. gregoryi shows a relatively undistorted piriform fenestra (fig. 21). This region is less well-preserved in UW 13508 (fig. 22), but also appears to exhibit an incompletely ossified tympanic roof.

Type Specimen:

KU 9112, left maxillary fragment with P3-M3, associated left dentary with p4-m3 (fig. 26).

Referred Specimens:

USNM 455680, skull (fig. 27) with left P3-M3, fragmentary right I2 and canine root, associated left (fig. 28) and still articulated right posterior mandibles with m1-m3 and m2-m3, respectively; DMNH 1747, skull (fig. 29) with left C-M2 and right canine; FMNH PM34512, rostrum with right and left P2, left P3 and partial P4, lingual roots of M1-M2, alveoli for right and left canines, associated with brain endocast, right lambdoid plate, and petrosal; and TMM 40492-9, rostrum with left P4-M2, right P4-M1, M3, partial P3, preserving pterygoid and orbitotemporal regions (fig. 30).

Temporal and Geographic Distribution:

Chadronian (late Eocene) of northeastern Colorado (Iliff 24, Fremont Butte 20) and central Wyoming (Beaver Divide 5); and late Duchesnean (late middle Eocene) of west Texas (Red Mound 32).

Diagnosis:

The type and referred specimens of Apternodus iliffensis are smaller than other species of Apternodus. They have reduced upper molar protocones that are continuous with the molar lingual cingula and a small M3. The anterior dentition of USNM 455680 and DMNH 1747 is premolariform, not bulbous; a diastema is absent between P2 and P3; and minute diastemata are present on either side of the canine. The lacrimal foramen is set off from the orbit by a prominent crest, formed by the proximity of the infraorbital canal. Medial to the promontory, the rostral tympanic process of the petrosal is flat. The external auditory meatus is weakly concave ventrally, and the ventral margin of the lambdoid plate does not extend far below the glenoid region, similar to the condition in A. baladontus.

Remarks:

Cranially, this species is the smallest within Apternodus, even though its cheek teeth are relatively larger than those of A. mediaevus and A. baladontus. Two nearly complete skulls are now known for A. iliffensis: USNM 455680 (figs. 27, 28) and DMNH 1747 (fig. 29). The lower dentition anterior to p4 remains unknown.

As indicated by DMNH 1747 and USNM 455680, the upper anterior dentition of A. iliffensis closely resembles that of A. brevirostris and A. gregoryi. That is, the P2, canine, and I2 are small, premolariform, non-bulbous teeth. I2 has one root, P2 and canine two. I2 has a small but distinct posterior cusp, connected to the apex of the tooth by a crest.

The basicranium is mediolaterally narrower than that of other species. However, structures of the posterior braincase (i.e., the lambdoid plate, temporal fossa, and sagittal and nuchal crests) are at least as robust as those of other taxa.

USNM 455680 displays a large, undistorted piriform fenestra anterior to the petrosal, bounded medially by a strut of alisphenoid, which defines the lateral border of the anterior carotid foramen. The smaller vidian foramen is present anteromedial to the anterior carotid foramen. The petrosal retains a subtle but distinct groove extending from the apex of the promontory, medial to the vestibular foramen, anteromedially toward the anterior carotid foramen. In life this groove was associated with the internal carotid artery and contradicts McDowell's (1958: 168) interpretation that the internal carotid artery of Apternodus had a medial course adjacent to the petrosal-basisphenoid suture.

Referring the west Texas specimen (TMM 40492-9; fig. 30) to A. iliffensis makes this species the most geographically widespread within Apternodus. Novacek (1976b) tentatively referred this specimen to “A. cf. brevirostris”. However, TMM 40492-9 contrasts with A. brevirostris, and is similar to the iliffensis type specimen, in having reduced molar protocones, a mesiodistally elongate P3, and a small M3. In fact, the M3 of TMM 40492-9 is considerably smaller than that of any other Apternodus specimen. It further differs from most other Apternodus individuals in showing a particularly elongate P3. Hence, at present we regard this specimen as a member of A. iliffensis; however, recovery of additional material from west Texas may justify the recognition of a new species.

In his 1978 paper describing a second apternodontid specimen (FMNH PM34512) from White River Formation exposures north of Iliff, Colorado, Galbreath made several confusing statements. First, Galbreath (1978: 302) stated that “there should not be any doubt that the new specimen [i.e., FMNH PM34512] is an apternodontid but not an apternotid.” However, Galbreath's intended meaning for the term “apternotid” is unclear. He may have been of the opinion that his new specimen should not be included in the genus Apternodus, in which case we would disagree for reasons given below.

Galbreath (1978: 301) indicated that “other details not mentioned in the original description of A. iliffensis acquire significance now with the discovery of a second apternodontid from the same locality”, and described some aspects of the A. iliffensis infraorbital canal, presumably based on FMNH PM34512. However, on the next page he argued that “the dissimilar details of the infraorbital canal suggests that the two [i.e., KU 9112 and FMNH PM34512] would not be in the same specific taxon”, in contrast to his previous implication that both specimens illuminate the anatomy of A. iliffensis.

Galbreath (1978: 299) stated that the “preorbital ridge so characteristic on the maxillaries of A. brevirostris and A. gregoryi” is missing on FMNH PM34512. However, a “preorbital ridge” has not been explicitly mentioned by other authors, and Galbreath does not illustrate or further elaborate on the identity of this character. There is some variation in the distance between the superior margin of the infraorbital canal to the lacrimal foramen; in A. brevirostris this distance is shorter than that in A. gregoryi and defines the border between the lacrimal foramen and orbit with a ridge. However, this feature serves to distinguish the two species, not characterize them. Furthermore, a ridge posterior to the lacrimal foramen is present in FMNH PM34512. Hence, we regard this character as a misinterpretation by Galbreath of breakage along the lateral margin of the maxilla in FMNH PM34512.

The fragmentary teeth of FMNH PM34512, consisting of an intact P2, P3, partial P4, and the lingual remnants of M1-M2, are identical in size and appearance to the corresponding parts of P3-M2 in KU 9112. Most importantly, both specimens lack protocones and are from temporally and geographically adjacent localities, and can easily be accommodated in the single species, Apternodus iliffensis.

Krishtalka and Setoguchi (1977) described upper and lower dental remains of “Apternodus sp. cf. A. iliffensis” from Badwater locality 20, now considered to be early Duchesnean in age. One of these specimens. CM 29012 (fig. 31), is a right maxillary fragment with P4-M1, not M1–2 as stated by Krishtalka and Setoguchi (1977), with weak protocones that presumably led these authors to associate the Badwater specimens with A. iliffensis. However, the protocones of CM 29012 are actually larger than those of the A. iliffensis type. Other potentially useful anatomical regions (e.g., P3 and M3) are missing. Hence, this specimen and others from Badwater locality 20 are not sufficiently diagnostic for identification at the species level. Nevertheless, the Badwater locality 20 Apternodus specimens, and similar material described by Storer (1995) from the approximately contemporaneous Lac Pelletier Lower Fauna of Saskatchewan, document the first occurrence of this genus in the middle Eocene of North America. The small “Apternodus sp.” described by Tong (1997) from China may be similar or even older in age.

Ostrander (1987) referred several isolated upper and lower teeth from the middle Chadronian Raben Ranch Local Fauna of Nebraska to A. iliffensis. However, Ostrander did not provide illustrations of these specimens, and the mesiodistal and buccolingual dimensions he reports are broadly consistent with those of several Apternodus species. In our experience, isolated upper or lower molars cannot be assigned with confidence to a single species of Apternodus.

Type Specimen:

UW 11046, nearly complete skull (fig. 32) preserving base of left I1, C, P3-M3, right I1, C-M3, missing posterior palate and pterygoid region, articulated right dentary with broken i1, c, broken p3, p4-m3, and associated left dentary (fig. 33) with p4-m3; associated ulna, proximal and distal fragments of right humerus (fig. 34), fragments of scapulae, cervical vertebrae, and ribs.

Referred Specimens:

UW 10981, rostrum associated with partial right dentary with p3-m3, broken i1, i3, c; UW 10984, fragmentary skull with left M1-M3, right M1-M2; UW 11291, partial rostrum with left P2-P4, broken M1-M2; UW 11292, fragmentary skull with right M1-M3, associated left maxilla with P3-M2; UW 11295, fragmentary rostrum with left P3-P4, articulated posterior left dentary, associated with anterior left dentary containing p4-m1; and UW 11296, rostrum with articulated left and right dentaries, with complete dentition except for I/i1 and P/p2.

Etymology:

The name “major” was coined originally by Kron (1978) and refers to the large size of this species.

Temporal and Geographic Distribution:

Chadronian (late Eocene) of eastern Wyoming (Dilts Ranch 10).

Diagnosis:

This species is the largest yet known in the genus Apternodus. It is anatomically distinctive by virtue of the laterally flared anterior lambdoid plates and the enlarged, bony torus present along the ventral margin of the external auditory meatus. Unlike A. iliffensis, the external auditory meatus is ventrally concave, but not to the extent seen in A. gregoryi (fig. 21). A. major also has a prominent, anteromedially running rostral tympanic process of the petrosal, a laterally prominent maxillary rudiment of the zygoma, and a lacrimal foramen continuous with the anterior orbit. Diastemata between the upper incisors and adjacent to P3 are lacking. The anterior dentition of A. major is large, but premolariform, resembling that of A. gregoryi. The large, two-rooted, premolariform P2 of A. major is similar in size to P3 and much larger than the P2 of A. brevirostris. Protocones on the upper molars are reduced and barely distinguishable from the lingual cingulum.

Remarks:

Several specimens collected and first described by Kron (1978) are well-preserved, but none is as complete as the type. This specimen (UW 11046) is a particularly large and robust individual, showing extreme development of the anteriorly flared lambdoid plates, a laterally expansive maxillary zygoma, and distinct fossae in the maxillae above each canine. A. major is found nowhere else except at the Dilts Ranch area examined by Kron (1978), which has also yielded other species of Apternodus (UW 13508 A. gregoryi and AMNH 74952 possibly A. brevirostris) but none of Oligoryctes. In addition to the details summarized in the diagnosis, the morphology of the type skull is as follows.

The maxilla adjacent to the anterior infraorbital canal is damaged on both sides. However, it can be seen on the left side to project far anterolaterally (fig. 32), providing extensive surface area for the attachment of anterior facial muscles. Following Butler (1956), the muscles that originate from this area include levator labii superioris, levator alae nasi, and zygomaticus. Whidden (ms in progress) has recently confirmed Butler's assessment for the origin of these muscles in insectivoran-grade taxa such as Solenodon. A. major possesses a relatively huge area for the attachment of anterior snout-muscles, somewhat more elaborate than that of Solenodon, which supports an elongate rod of cartilage that stretches anteriorly from the external nares. Hence, the rostral morphology of A. major, and that of Apternodus generally, appears compatible with an anteriorly elongate, cartilaginous extension of the proboscis (as reconstructed anterior to the elongate upper incisor in the frontispiece of this monograph).

The dorsal braincase of A. major is rugose (fig. 32). Due in part to its laterally extensive anterior lambdoid plate, A. major has more surface area for the attachment of the temporalis muscle than other species of Apternodus. The robust temporalis musculature of A. major is also reflected in the pronounced muscle scar on the lateral aspect of the mandibular coronoid process (fig. 33).

As in other species of Apternodus, A. major shows prominent piriform fenestrae in the tympanic roof anterior to the pars cochlearis of the petrosal. Anteromedial to the promontory, on the lateral margin of the basisphenoid, a distinct vidian foramen is evident. The anterior carotid foramen appears to be confluent with the piriform fenestra (fig. 32). Anterior and slightly ventral to the caudal tympanic process of the right petrosal, UW 11046 preserves fragments of what might be a broken ectotympanic or tympanohyal (fig. 32).

The posterior palate and pterygoid region are not preserved in UW 11046, but from other, more fragmentary specimens (e.g., UW 10981, UW 10984, UW 11292) they can be inferred to be similar to those of other species of Apternodus. The jaw joint, well preserved in the type, also resembles that of other Apternodus, as described above for A. baladontus.

UW 11046 is one of the few Apternodus specimens with associated postcrania. These include a left ulna and both proximal and distal left humeral fragments (fig. 34). Both elements resemble those of Solenodon, but are smaller. The proximal humerus shows a broad greater tuberosity and well-defined fossa for the biceps tendon. The distal humerus shows an entepicondylar foramen, prominent medial epicondyle, well-defined trochlea and capitulum, and a small fossa posterior to the capitulum that is continuous with the supinator crest. The proximal ulna differs from that of Solenodon in having a smaller, more gracile olecranon process.

Distal fragments of both scapulae, a clavicle, vertebrae, and ribs are also preserved together in an associated block of matrix. The right scapula appears to be articulated with the proximal end of the right clavicle. On both sides, the scapulae are unremarkable, with concave glenoid fossae for articulation with the proximal humeri. The neural arches of several vertebrae, probably thoracic, are partially exposed. Spinous processes of these vertebrae were either very small or not preserved in this specimen.

Type and Only Specimen:

UW 14072, crushed but otherwise complete skull (fig. 35) preserving left I1, C-M3, right I1, broken C, P2, broken P3, P4-M3, articulated left dentary with complete dentition, associated right dentary with i1-i2, c-m3 (fig. 36), associated tibia, atlas, axis, partial scapula, proximal humerus, distal ulna, ribs, and other fragmentary postcranial fragments (fig. 37). All elements are slightly distorted due to postmortem deformation.

Etymology:

This species is named in honor of the faculty, students, and staff of the Department of Geology and Geophysics, University of Wyoming, under whose auspices the type and only specimen was collected and is now curated. The Greek word transliterated as “dasophylakas” is an interpretation of the English word “cowboy”, the UW mascot. Literally, the trivial name means “ranger” or “guardian of wild land and forest” (N. Soulunias and I. Trivilas, personal commun.).

Temporal and Geographic Distribution:

Chadronian (late Eocene) of southeastern Wyoming (Harshman Quarry 21).

Diagnosis:

Protocones on the upper molars are absent, a condition somewhat exacerbated by wear. The anterior dentition is premolariform, not bulbous. P3 has an elongate stylar margin, and diastemata are evident on either side of the upper canine, between I1 and I2, and probably also between P2 and P3; some distortion in the maxilla between P2-P3 on both sides precludes certainty as to the size of the P2-P3 diastema. Both P2 and p2 are smaller than adjacent teeth and are similar in proportion to those of A. brevirostris. The mental foramen on the dentary is located just below p2, a position slightly anterior to that of other Apternodus species. The p3 shows a slight buccal cingulid and a distinct posterior cusp. The external auditory meatus of UW 14072 is not marked by a torus, in contrast to that of A. major, but is flat as in A. iliffensis. Another similarity to A. iliffensis is the ventral margin of the lambdoid plate, which reaches just slightly ventral to the level of the jaw joint. The rostral tympanic process of the petrosal is flat. A piriform fenestra is evident, lateral to a small anterior carotid foramen.

Remarks:

UW 14072 is a large version of A. iliffensis, sharing with it aspects of the ear and lambdoid region as well as reduced upper molar protocones and an elongate P3 (polymorphic in A. iliffensis). UW 14072 differs qualitatively from individuals of A. iliffensis in having an M3 similar in buccolingual width to the M2, an elongate rostrum with diastemata on either side of the canine, and a transversely wide basicranium. It differs from other Apternodus species in having a mental foramen ventral to p2; the position of the mental foramen(ina) is unknown in A. iliffensis. Cranial proportions are similar to those of A. gregoryi, with the exception that A. dasophylakas has a very shallow external auditory meatus and a lambdoid plate that does not extend far ventral to the jaw joint. The orbitotemporal region of UW 14072 is crushed, leaving a very distorted lacrimal foramen on the left side, and none on the right. It is therefore difficult to tell if the posterior border of the lacrimal foramen is flush with the anterior orbit as in Apternodus gregoryi (figs. 21, 22) and A. major (fig. 32). The left craniomandibular joint, with the jaw still in articulation, preserves what might be part of an ossified articular disk separating the condyle from the glenoid fossa (fig. 35). However, due to the postmortem distortion of the specimen, it is difficult to discount the possibility that this region contains multiple fragments of the mandibular condyle itself.

Postcranially, what little is known of A. dasophylakas (fig. 37) does not appear to be specialized for a nonterrestrial locomotor repertoire. The only exception to this observation in any Apternodus specimen may be the structure of the proximal femur in A. gregoryi (UW 13508; see above). The distal tibia of A. dasophylakas is not fused with the fibula. The scapula has a triangular shape, narrow laterally at the glenoid fossa and broadening dorsomedially toward the vertebral column. The second cervical vertebra shows a ventrally projecting keel on the posteroventral aspect of its centrum (similar to the axis of shrews and Echinosorex); and the first cervical vertebra shows two widely separated facets for the occipital condyles of the skull.

Included Genera:

Oligoryctes Hough, 1956

Distribution and Diagnosis:

As for Oligoryctes.

Type Species:

Oligoryctes cameronensis Hough, 1956

Included Species:

O. altitalonidus Clark, 1937 (new combination) and the Tabernacle Butte taxon (unnamed in this paper; see below).

Temporal and Geographic Distribution:

Bridgerian (late early Eocene) through Orellan (early Oligocene) throughout the North American western interior, including Wyoming, Montana, Colorado, North Dakota, South Dakota, Saskatchewan, California, Nevada, Utah, Texas, and Nebraska, possibly including Asia if material described by Tong (1997) is included.

Diagnosis:

Oligoryctes is a shrew-sized animal with zalambdodont molars, lacking metacones and with reduced talonid basins. The m3 talonid cusp is slightly taller than the m3 paraconid. The upper molars have distinct protocones and anterior cingula, and the posterior two lower incisors are tricuspid. The medial aspect of the coronoid process is deep or pocketed. Oligoryctes has a relatively unspecialized posterior braincase without the elaborate lateral extensions of the squamosal, petromastoid, and occipital that form the lambdoid plates in Apternodus. Instead, Oligoryctes has a laterally rounded squamosal and petromastoid. It shows a prominent entoglenoid process for posterior support of the jaw joint, medial to the postglenoid foramen and anterior to the promontory of the middle ear. The basicranium shows an enlarged foramen ovale and lacks an alisphenoid canal. The squamosal extends posteriorly along the ventrolateral margin of the braincase, lateral to the large piriform fenestra. The anterior exit of the sinus canal and ethmoid foramen are located well anterior to the sphenorbital fissure and are not incorporated into its superior margin. As in Apternodus, the lacrimal foramen is large and laterally oriented.

Type Specimen:

USNM 19909, rostrum (38) with right P2-P4, broken M1, M2-M3, left P3-M3, preserving pterygoid region, associated left dentary with p3-m3 (fig. 39); original lost, known only from illustrations.

Referred Material:

USNM 516840, undistorted cranium (fig. 40), missing premaxillae and mandibles, preserving left ?C, P2, dP3, erupting P3, P4-M3, right P2, dP3, erupting P3, P4-M3; USNM 516846, left dentary with p3-m3; UCM 52446, rostrum (fig. 41) with left broken I2, I3-M3, right I2, P2-M3, orbitotemporal region, articulated left dentary with complete dentition except for missing crown of i3, associated right dentary with p3-m3 (fig. 42); MPUM 6677, left dentary with heavily worn i1-m3; MPUM 6859, right ?M1, left maxillary fragments with P3-M3, associated left petromastoid fragment preserving superior semicircular canal and stapedius fossa, missing bulk of pars cochlearis; CM 17193, rostrum with right and left P3-M3 plus anterior alveoli (fig. 43; original lost, known only from photos); CM 73977, associated right and left dentaries, each with p4-m3.

Temporal and Geographic Distribution:

Chadronian (late Eocene) of central Wyoming (Cameron Spring 5, Flagstaff Rim 18) and western Montana (Pipestone Springs 29, Eureka Valley Rd. 16), probably also northwestern Nebraska (33; see Ostrander, 1987).

Diagnosis:

The small maxillary rudiment of the zygoma is located lateral to the posterior edge of M1. The type skull shows a foramen in the alisphenoid posterior to the enlarged foramen ovale, similar in position to the alisphenoid foramen transmitting the inferior stapedial ramus in some tenrecs (Asher, 2001). A small foramen is evident on the rostrum near the nasal-frontal suture, as are foramina laterally along the path of the sinus canal, above the external auditory meatus and jaw joint. The dental formula is 3.1.3.3/3.1.3.3, differing from that of Apternodus in having one more upper incisor. The crown of P3 is longer than that of P4; both show a prominent, anteriorly projecting parastyle. The trigonids of the lower molars decrease progressively in width, with that of m1 the largest and m3 the smallest. The talonid of the m3 is elongate, making this tooth the longest in the molar series. As in O. altitalonidus, the m3 talonid cusp is slightly greater in height than the m3 paraconid; the medial aspect of the coronoid process is pocketed; and the coronoid process is externally convex.

Remarks:

USNM 516840, collected by one of us (RJE) from Flagstaff Rim, is by far the best specimen yet known of this taxon (fig. 40). It is one of the few small “apternodontid” specimens known that preserves petrosals and elements of the deciduous dentition; and its pristine state of preservation is also quite remarkable. The undistorted rostrum shows sutures between the nasals and maxilla; the nasal-frontal suture is faint, but appears to delimit a posteriorly narrow nasal. A small foramen is present in the rostrum at the posterior margin of the nasal. Premaxillae are missing. The lacrimal foramen is large, laterally directed, and situated on the anterior aspect of the infraorbital canal. The frontal-maxillary suture is also visible and appears to extend well into the orbit, accompanied by the roots of the maxillary teeth (fig. 40). The maxilla anterior to the infraorbital canal is also perforate and shows roots of the erupting permanent P3. The posterior ethmoid region is broad and contributes to a slight postorbital constriction.

The braincase is well preserved and shows no sign of sutures within the frontal or between the frontal, parietal, and occipital, although numerous hairline cracks, some bilateral, may represent such sutures. In contrast, the petromastoid shows well-defined sutures with the occipital, parietal, and squamosal, and is well-exposed laterally. The petromastoid-squamosal suture extends ventrally onto the basicranium and defines the lateral boundary of the tympanic cavity. The squamosal also shows fairly well-defined sutures with the parietal and alisphenoid. On the squamosal side of the squamosal-parietal suture are three foramina, two located adjacent to the petromastoid, and one dorsal to the jaw joint.

USNM 516840 shows four foramina anterior to the sphenorbital fissure. The anteriormost is a large ethmoid foramen, providing a conduit into the anterior cranial fossa and posterior ethmoid region. Immediately posterior and ventral is a much smaller foramen which may also connect the orbit with the anterior cranial fossa. Continuing posteriorly, the next large foramen is the anterior exit point of the sinus canal; slightly medial to that is a small optic foramen. Both the sinus canal and the optic foramen are located immediately anterior to the large sphenorbital fissure, the ventral boundary of which is continuous with the pterygoid. Posterior to the palate, sutures are evident between the V-shaped vomer and pterygoids.

The most conspicuous openings on the basicranium of USNM 516840 are the large piriform fenestra and foramen ovale. The latter is slightly larger than the sphenorbital fissure, and the piriform fenestra is as large as the bony promontory itself. Posterior to foramen ovale, and within the base of the entoglenoid process, is a foramen that served as a conduit for the inferior ramus of the stapedial artery (mentioned above as similar in position to that of tenrecs such as Geogale; see Asher, 2001). Subtle grooves on the promontory bone indicate that the internal carotid artery entered the tympanic cavity at the ventral apex of the pars cochlearis, just medial to the prominent caudal tympanic process of the promontory. It immediately bifurcated into a proximal stapedial ramus, which traversed laterally toward the fenestra vestibuli, and a transpromontorial internal carotid (sensu Wible, 1986), which ran anteromedially, entering the braincase via the piriform fenestra. A foramen in the basisphenoid at the dorsomedial margin of the tympanic cavity is smaller in caliber than the groove for the internal carotid on the pars cochlearis. Hence, we believe this foramen provided a conduit for a small vidian artery, branching from the internal carotid immediately proximal to the latter's path through the piriform fenestra.

Zygomatic arches are completely lacking, and both maxillary and squamosal rudiments of the zygoma are reduced. As in Apternodus, the jaw joint is defined posteriorly by the entoglenoid, not postglenoid, process. This is located medial to the postglenoid foramen and anterior to the middle ear.

The upper molars of USNM 516840 are fully erupted and show minimal wear. P4 is almost fully erupted; P3 is still in its crypt, but some cusps are evident dorsal to dP3. P2 is fully erupted and minute in size. The right upper canine is missing; on the left side it is intact but not yet fully erupted and is displaced posteriorly.

Also noteworthy among the new specimens of O. cameronensis is UCM 52446, a well-preserved rostrum (fig. 41) and associated, largely complete mandible (fig. 42). Its dentition is fully erupted and shows a permanent P3 that is larger than the dP3 of USNM 516480, particularly the paracrista running posterior to the paracone. The dentary shows trigonids of m1-m3 becoming progressively smaller posteriorly, as well as multicuspid incisors. Otherwise, UCM 52446 conforms with the morphology evident in USNM 516840.

Some of Hough's (1956) original description of Oligoryctes does not accord with the material that we now have available. Unfortunately, the type specimen designated by her (USNM 19909) now appears to be lost; however, one of us (MCM) previously had an opportunity to examine it. Her diagnosis stated (1956: 538) that the basicranial region differs from that of Apternodus “in having lateral descending processes of the basisphenoid, which embrace medially the small anterior bullae.” The features in question are actually the medial walls of the sphenorbital fissures; and the contents of the “anterior bullae” consist principally of matrix fillings of the trigeminal nerve tracts. If the lower teeth are placed in occlusion with the uppers, this area may be seen to lie well in front of the jaw suspensorium. The medially excavated coronoid process of the mandible was not seen by Hough because the specimen was incompletely prepared at the time it was described.

McDowell (1958: 171, 172) also mistakenly regarded Oligoryctes as possessing an ossified auditory bulla, going even farther than Hough by arguing (p. 171) that Apternodus also possessed an ossified bulla, based on a presumed close relationship between the two genera. However, elsewhere in his paper (McDowell, 1958: 167–168), he seems to have recognized Hough's mistake: “Hough (1956) claims that the basisphenoid tympanic wing is present in Oligoryctes, a genus probably closely related to Apternodus. However, this is not clear from the specimen and appears rather to be but a slight selvage of basisphenoid against the ear chamber and does not have the characteristic position of the lipotyphlan basisphenoid, abutting against the Eustacian cartilage.” In any event, new material described here of both Apternodus and Oligoryctes clearly demonstrates that the middle ear of both taxa is similar to that of soricids and Solenodon, with neither taxon showing an ossified auditory bulla, and both showing a piriform fenestra (polymorphic within A. brevirostris and A. mediaevus).

Type Specimen:

YPM PU13774, left dentary with p4-m3; original lost but casts remain (fig. 44).

Referred Material:

USNM 516843, nearly complete skull (fig. 45), slightly compressed dorsoventrally, missing petrosals, with associated mandibles with left i1-m3 and right p3-m3 (fig. 42); USNM 22816, associated right maxillary fragments with P2-M3 (fig. 46); USNM 516841, left dentary with m1-m3; USNM 516842, rostrum with nearly complete dentition including dP3 and dP4 and erupting left P3-P4 in crypts, plus associated mandibles with complete dentition; USNM 516847, left dentary with p3-m3; USNM 516848, left dentary with p4-m3; USNM 516849, left dentary with p4-m3; USNM 516850, right dentary with p3-m3; USNM 516851, right dentary with p4-m3; USNM 516852 left dentary with m1; USNM 516853, left dentary with p3-m3; USNM 516854, rostrum with right and left P3-M3; USNM 516855, rostrum with heavily worn left C-M3 and right P3-M3; USNM 516856, rostrum with right C-M3 and left C-P4; USNM 516857, rostrum with right P3-M3 and left P3-P4; USNM 516858, right dentary with m2-m3; USNM 516862, left maxilla with P3-M2; USNM 516863, right maxilla with P3, M1-M3; USNM 516864, right maxilla with M1-M3; USNM 516865, left dentary with m1-m3; USNM 516866 left maxilla with M1-M2; USNM 516867, anterior braincase and rostrum with right P3-M3 and left P4-M3; USNM 516868, left dentary with m1-m3; USNM 516869, left dentary with p3-m3; USNM 516870, left dentary with m2; AMNH 105310, right dentary with p4-m3; MPUM 0414, left dentary with p3-m3; MPUM 2592, left dentary with m2-m3; MPUM 6797, left dentary with p4-m2; MPUM 6857, left dentary with p4-m3; MPUM 6858, rostrum with right and left P3s; CM 9260, right dentary with p3–4, m2–3; CM 71571, left dentary with p4-m2; CM 71572, left dentary with m1-m2; CM 71573, left dentary with m2–3; and CM 73976, right maxilla with P4-M2.

Temporal and Geographic Distribution:

Uintan (middle Eocene) through Orellan (early Oligocene) of Wyoming (Flagstaff Rim 18, East Fork Basin 12), Montana (Pipestone Springs 29, Diamond O Ranch 9, Little Pipestone Creek 25, Cook Ranch 7), South Dakota (Big Badlands 4), and North Dakota (Fitterer Ranch 17). Geographic distribution probably also includes Saskatchewan (8; see Storer, 1996) and California (34; see Walsh, 1996).

Diagnosis:

Unlike Apternodus, and like most other mammals including Oligoryctes cameronensis, O. altitalonidus lacks the extensive, box-shaped structures of the posterior braincase. Instead, its braincase is rounded and gracile, as in Microgale. O. altitalonidus possesses a maxillary rudiment of the zygoma lateral to M2, slightly posterior to that of O. cameronensis. O. altitalonidus is smaller than all other Tertiary zalambdodonts except for the unnamed Tabernacle Butte taxon (McKenna et al., 1962; Bloch et al. in prep.). Perhaps the most conspicuous similarity with O. cameronensis is the large size of foramen ovale, which is greater in diameter than the jugular foramen (fig. 45). The squamosal is posteriorly elongate, forming the lateral border of the large piriform fenestra. The stylar crests of P3 are elongate, as in O. cameronensis. In contrast to Apternodus and O. cameronensis, this species possesses a p1, one of eight teeth anterior to the three lower molars; hence, its dental formula is 3.1.3.3/3.1.4.3. The p3 shows a cingulid buccally. As in O. cameronensis, the m3 is longer than more anterior cheek teeth, and the posterior two incisors are tricuspid. The tall trigonid cusp on m3, for which this species was named, is also present in the larger O. cameronensis and the Tabernacle Butte taxon. Unlike O. cameronensis, the molar trigonids of O. altitalonidus are similar in size and do not decrease markedly from m1 to m3. The medial side of the coronoid process is pocketed, as in modern soricids. Unlike soricids, the coronoid process is bowed in shape, or externally convex.

Remarks:

USNM 516843 (figs. 42, 45) is the best specimen of O. altitalonidus yet known. Several additional specimens from Pipestone Springs (USNM 22816; fig. 46) and Flagstaff Rim (e.g., USNM 516842, 516854, and 516857) confirm details of cranial anatomy of this genus evident from USNM 516843. USNM specimens from Fitterer Ranch (516868–516870) document the previously unknown presence of this species in the Orellan of North Dakota. In addition to fully erupted M1-M3, USNM 516842 shows an erupting P3 and P4 in the process of replacing their deciduous precursors, demonstrating that unlike shrews, O. altitalonidus had functional deciduous teeth. As in O. cameronensis, P4 erupts prior to P3.

Oligoryctes altitalonidus resembles O. cameronensis more than it does any species of Apternodus. It shares a zalambdodont dentition with both Apternodus and O. cameronensis, but lacks the derived shape of the Apternodus posterior braincase. O. altitalonidus shares with O. cameronensis an enlarged foramen ovale, an elongate P3, and multicuspid lower incisors. For these reasons, many previous authors (e.g.,. Emry, 1979; Tabrum et al., 1996; Robinson and Kron, 1998) have referred to this species using the name Oligoryctes altitalonidus, a combination that we support.

Type Specimen:

To be determined by Bloch et al. (manuscript in progress).

Referred Material:

USNM 417465, left maxillary fragment with M1-M3; USNM 417464, distorted left dentary with p3-m3; CM 13627, left dentary with m1-m3, original lost but casts remain (see fig. 47); CM 13859, left dentary with m2; DMNH 8776, right dentary fragment with broken m1, talonid of m2, and partial alveolus for m3; probably also AMNH 55689, edentulous mandibular fragment mentioned by Simpson in McGrew et al. (1959).

Temporal and Geographic Distribution:

Bridgerian (early middle Eocene) of eastern Nevada (Elderberry Canyon 14), southwestern Wyoming (Tabernacle Butte 37), northeastern Utah (Powder Wash 31), and Uintan (middle Eocene) of northwestern Colorado (Sand Wash 35).

Remarks:

Among taxa that lack a talonid basin and metacone, the Tabernacle Butte species is unique by virtue of its deep anterior cingulum and large protocones of the upper molars, medially deep but not fully pocketed coronoid process, and extreme small size.

Several authors, including McKenna et al. (1962) and Emry (1990), have acknowledged the presence of this as yet unnamed, diminutive, early-middle Eocene “apternodontid”. In fact, based on previously circulated, unpublished versions of this paper, Romer (1966) prematurely called this taxon “Eoryctes”, a name now unavailable as it was used by Thewissen and Gingerich (1989) for a genus of palaeoryctid. As of early 2002, J. Bloch et al. have a manuscript in progress that will name this taxon, including additional material from southwestern Wyoming in the University of Michigan collections. We defer to them for nomenclature, but will use available material of this species as an informally named terminal taxon (i.e., “Tabernacle Butte taxon”) in our phylogenetic analyses.

Included Genera:

Parapternodus Bown and Schankler, 1982; Koniaryctes Robinson and Kron, 1998.

Temporal and Geographic Distribution:

Wasatchian (early Eocene) of northern Wyoming (Clark's Fork, Bighorn, and Powder River basins).

Diagnosis:

Known only from fragmentary teeth and jaws, the members of this small, shrew-sized family have zalambdodont molars with a reduced, unbasined talonid and lack buccal cingulids and an m3 posterior cusp. The m3 is anteroposteriorly shorter than m1 or m2. The anterior incisor is enlarged, leaving a large alveolus that extends posteriorly along the base of the jaw at least as far posteriorly as p3 (figs. 48, 49).

Type and Only Species:

P. antiquus.

Distribution and Diagnosis:

As for P. antiquus.

Type Specimen:

YPM 31169, left dentary with m2–3.

Referred Material:

UMMP 81557, right dentary with p4-m2, m2 trigonid broken; UMMP 81560, right dentary with m2-m3 and base of coronoid process (fig. 48); UMMP 81561, left dentary with a fragmentary incisor root and p4-m3 (fig. 49); UMMP 81558, maxillary fragment with left ?P3; UMMP 81559, left dentary fragment with lower molar; UMMP 81562, maxillary fragment with right upper molar; UMMP 81563, maxillary fragment with left P4 or M1.

Temporal and Geographic Distribution:

Wasatchian (early Eocene) of northern Wyoming (Clark's Fork Basin 19).

Diagnosis:

The talonids of Parapternodus are exceedingly small and form no basin. Anterior to m3, the talonids consist only of a ridge capped by a weak cusp. Parapternodus is slightly smaller than O. cameronensis (contra Bown and Schankler, 1982) and is similar in size to O. altitalonidus. It differs from the latter in having a premolariform p4, an anteroposteriorly short m3 talonid, and mental foramina adjacent to each p4 root. It also lacks buccal cingulids on its lower molars. UMMP 81561 preserves a fragmentary, enlarged root of an anterior incisor that extends posteriorly at least as far as the region below p3. Although no specimen retains an intact coronoid process, the region immediately posterior to m3 on UMMP 81560 and 81561 is not expanded transversely, indicating that the coronoid process is more gracile than that of Apternodus. The base of the coronoid process in UMMP 81650 is excavated deep to the alveolar plane of the dentary, suggesting that like soricids and the larger two species of Oligoryctes, Parapternodus also possessed a coronoid process that was pocketed medially. Alternatively, this excavation could be a dorsally broken entrance for the inferior alveolar nerve and artery; however, we consider this explanation unlikely because of the considerable dorsoventral size of this space and its anterodorsal proximity to the toothrow (fig. 48).

Remarks:

Perhaps due to the fragmentary type material, Bown and Schankler (1982: 67) used some continuous, unquantified comparisons to define Parapternodus, such as “m2 paracristid and postvallid relatively narrower transversely than in Oligoryctes or Apternodus …” and “trigonid less compressed anteroposteriorly than Oligoryctes”. These statements cannot be verified based on presently available material. Nevertheless, Bown and Schankler (1982) accurately identified a valid species based on an extremely limited sample, which has improved only slightly since 1982.

Type and Only Species:

Koniaryctes paulus Robinson and Kron, 1998

Distribution and Diagnosis:

As for K. paulus.

Type Specimen:

UCM 58291, right dentary fragment with m1-m2.

Referred Material:

UCM 59843, right dentary fragment with broken m2, preserving talonid (fig. 50).

Temporal and Geographic Distribution:

Wasatchian (early Eocene) of northeastern Wyoming (Powder River Basin 30).

Diagnosis:

Koniaryctes is slightly larger than Parapternodus. As in other zalambdodonts, the talonid is reduced to a crest and has no basin. K. paulus lacks buccal cingulids and has an anteroposteriorly short m3. In contrast to Parapternodus, there is no evidence of a notch below the paraconid into which the talonid rudiment of the anterior tooth would fit. This is consistent with Robinson and Kron's (1998) observation that the talonid cusp is absent throughout the molar toothrow. A second specimen (UCM 59843) preserves an unbasined m2 talonid which lacks a cusp, and is therefore referable to Koniaryctes. The rudiment of an enlarged alveolus for an anterior incisor is evident below the p4 alveoli; this specimen also preserves alveoli for m1, m3, the posterior root of p3, and a mental foramen below p4. The dentary lateral to m3 shows a breakage scar that was probably connected to the base of a robust, anteriorly placed coronoid process (fig. 50).

Remarks:

The position of the mental foramen below p4 in UCM 59843 supports the identification of the type specimen (UCM 59821) as m1-m2; the dentary of the latter specimen shows no mental foramina. Without this association, the possibility that the type specimen represents p4-m1, and not m1-m2 as stated by Robinson and Kron (1998), would be difficult to reject. The absence of a talonid cusp and paraconid notch support the status of Koniaryctes paulus as a real species, distinct from Parapternodus antiquus. Moreover, if UCM 59843 is correctly attributed to it, Koniaryctes is further distinguishable from Parapternodus (and is similar to Apternodus) by virtue of a robust, anteriorly extensive coronoid process. Nevertheless, this sample is even smaller than that of Parapternodus, and a definitive evaluation of its validity must await the discovery of better material.

Type Specimen:

To be determined by C. B. Wood (manuscript in progress).

Referred Specimens:

YPM PU16520, endocast with partial edentulous skull; YPM PU16521, edentulous skull missing rostrum, basicranium intact (fig. 51).

Temporal and Geographic Distribution:

Puercan (early Paleocene) and Tiffanian (late Paleocene) of Park County, northwestern Wyoming (Ft. Union Formation 19).

Remarks:

YPM PU16521 has laterally expansive, boxlike muscle attachments on its posterior braincase, reminiscent of those in Apternodus. The same is presumably true for YPM PU16520, although we have not been able to observe this specimen directly. Based on the appearance of the posterior braincase in these specimens, Edinger (1964) and Sloan (1969) suggested that Apternodus was present in the Paleocene. Unlike Apternodus and Oligoryctes, YPM PU16521 has a true postglenoid process supporting the jaw joint posteriorly (fig. 51). As in many placental mammals (e.g., Leptictis; see McDowell, 1958: fig. 25), but not most insectivorans, this process is located lateral to the postglenoid foramen and middle ear. An entoglenoid process is present, but does not support the mandibular condyle posteriorly. YPM PU16521 also shows remnants of an ossified auditory bulla, comprised at least in part by the rostral tympanic process of the petrosal and basisphenoid, and possibly also by an entotympanic.